Systems and methods for matching face shapes

Abstract

Systems for matching face shapes may include a computer-readable non-transitory storage medium and an executing hardware unit. The storage medium may include a set of instructions for target object shape matching. The executing hardware unit may be in communication with the storage medium and may be configured to execute the set of instructions. The executing hardware unit may be configured to obtain a target object image for shape matching; determine a shape character of the target object image based on a shape of the target object image; determine similarities between the target object image and a plurality of template images of reference objects based on the shape character of the target object image and shape characters of the reference objects in the plurality of template images; and select a template image from the plurality of template images that has a largest similarity to the target object image.

What is claimed is: 
     
       1. A system for shape matching, comprising:
 at least one computer-readable non-transitory storage medium, comprising a set of instructions for target object shape matching; and 
 at least one executing hardware unit in communication with the at least one computer-readable non-transitory storage medium that is configured to execute the set of instructions and is configured to:
 obtain a target object image for shape matching; 
 determine a shape character of the target object image based on a shape of the target object image; 
 determine similarities between the target object image and a plurality of template images of reference objects based on the shape character of the target object image and shape characters associated with the plurality of template images; and 
 select from the plurality of template images a template image that has a largest similarity to the target object image; 
 
 wherein each of the plurality of template image is an image of a reference object and is associated with a shape character of the image of the reference object; 
 wherein the shape character of the image of the reference object comprises coordinates of a plurality of discrete points locating along edges of components on the image of the reference object and a plurality of geometric blurring descriptors of a shape of the reference object; and 
 wherein the shape character of the target object image comprises coordinates of the plurality of discrete points locating along edges of components on the target object image and a plurality of geometric blurring descriptors of the shape of the target object. 
 
     
     
       2. The system of  claim 1 , wherein the at least one executing hardware unit is further configured to:
 obtain a template image of the plurality of template images; 
 obtain a standardized template image from the template images; and 
 determine the shape character of the image of the reference object based on the standardized template image. 
 
     
     
       3. The system of  claim 2 , wherein to obtain the shape character of the image of the reference object, the at least one executing hardware unit is further configured to:
 obtain a main shape of the reference object in the standardized template image; and 
 determine the shape character of the main shape, 
 wherein the main shape is a predetermined portion of the reference object. 
 
     
     
       4. The system of  claim 2 , wherein each of the standardized template images is categorized as at least one of a pet class image and a food class image. 
     
     
       5. The system of  claim 3 , wherein to obtain the main shape of the standardized template image, the at least one executing hardware unit is further configured to:
 apply an Active Shape Model Algorithm on the standardized template image when the reference object is an animal. 
 
     
     
       6. The system of  claim 3 , wherein to obtain the main shape of the standardized template image, the at least one executing hardware unit is further configured to:
 apply a Snake Model Algorithm on the standardized template image to position an outer boundary of the image of the reference object when the reference object is food. 
 
     
     
       7. The system of  claim 3 , wherein when the standardized template image is an image with a fuzzy contour to obtain the main shape of the standardized template image, the at least one executing hardware unit is further configured to:
 apply a Canny edge detection method on the standardized template image to obtain overall edges of the image of the reference object; and 
 filter out noises on the overall edges. 
 
     
     
       8. The system of  claim 3 , wherein to determine the shape character of the template image the at least one executing hardware unit is further configured to:
 obtain a blurred shape corresponding to the template image by conducting geometric blurring on the main shape of the image of the reference object in the template image; 
 determine the shape character of the template image by conducting sampling on the blurred shape corresponding to the template image. 
 
     
     
       9. The system of  claim 1 , wherein to determine the similarity between the target object image and a template image of the plurality of template images, the at least one executing hardware unit is further configured to:
 obtain a first difference that reflects a difference between a blurring operator of the shape character of the target object image and a blurring operator of the shape character of the template image according to the geometric blurring descriptors of the shape character of the target object image and the geometric blurring descriptors of the shape character of the template image; 
 obtain a second difference that reflects a difference between the shape character of the target object image and the shape character of the template image according to the coordinates of the plurality of discrete points of the shape character of the target object image and the coordinates of the plurality of discrete points of the template image; and 
 obtain the similarity between the shape character of the target object and the shape character of the template image according to the first difference and the second difference. 
 
     
     
       10. A computer-implemented method for shape matching, comprising:
 obtaining, by at least one computer hardware unit, a target object image for shape matching; 
 determining, by at least one computer hardware unit, a shape character of the target object image based on a shape of the target object image; 
 determining, by at least one computer hardware unit, similarities between the target object image and a plurality of template images of reference objects based on the shape character of the target object image and shape characters associated with the plurality of template images; and 
 selecting, by at least one computer hardware unit, from the plurality of template images a template image that has a largest similarity to the shape of the target object image; 
 wherein each of the plurality of template image is an image of a reference object and is associated with a shape character of the image of the reference object; 
 wherein the shape character of the image of the reference object comprises coordinates of a plurality of discrete points locating along edges of components on the image of the reference object and a plurality of geometric blurring descriptors of a shape of the reference object; and 
 wherein the shape character of the target object image comprises coordinates of the plurality of discrete points locating along edges of components on the target object image and a plurality of geometric blurring descriptors of the shape of the target object. 
 
     
     
       11. The computer-implemented method of  claim 10 , further comprising:
 obtaining, by at least one computer hardware unit, a template image of the plurality of template images; 
 obtaining, by at least one computer hardware unit, a standardized template image from the template images; and 
 determining, by at least one computer hardware unit, the shape character of the image of the reference object based on the standardized template image. 
 
     
     
       12. The computer-implemented method of  claim 11 , wherein obtaining the shape character of the image of the reference object comprising:
 obtaining, by at least one computer hardware unit, a main shape of the reference object in the standardized template image; and 
 determining, by at least one computer hardware unit, the shape character of the main shape, 
 wherein the main shape is a predetermined portion of the reference object. 
 
     
     
       13. The computer-implemented method of  claim 11 , wherein each of the standardized template images is categorized as at least one of a pet class image and a food class image. 
     
     
       14. The computer-implemented method of  claim 12 , wherein obtaining the main shape of the standardized template image comprising:
 applying, by at least one computer hardware unit, an Active Shape Model Algorithm on the standardized template image when the reference object is an animal. 
 
     
     
       15. The computer-implemented method of  claim 12 , wherein obtaining the main shape of the standardized template image comprising:
 applying, by at least one computer hardware unit, a Snake Model Algorithm on the standardized template image to position an outer boundary of the image of the reference object when the reference object is food. 
 
     
     
       16. The computer-implemented method of  claim 12 , wherein when the standardized template image is an image with a fuzzy contour obtaining the main shape of the standardized template image comprising:
 applying, by at least one computer hardware unit, a Canny edge detection method on the standardized template image to obtain overall edges of the image of the reference object; and 
 filtering out noises on the overall edges. 
 
     
     
       17. The computer-implemented method of  claim 12 , wherein determining the shape character of the template image comprising:
 obtaining, by at least one computer hardware unit, a blurred shape corresponding to the template image by conducting geometric blurring on the main shape of the image of the reference object in the template image; 
 determining, by at least one computer hardware unit, the shape character of the template image by conducting sampling on the blurred shape corresponding to the template image. 
 
     
     
       18. The computer-implemented method of  claim 10 , wherein determining the similarity between the target object image and a template image of the plurality of template images comprising:
 obtaining, by at least one computer hardware unit, a first difference that reflects a difference between a blurring operator of the shape character of the target object image and a blurring operator of the shape character of the template image according to the geometric blurring descriptors of the shape character of the target object image and the geometric blurring descriptors of the shape character of the template image; 
 obtaining, by at least one computer hardware unit, a second difference that reflects a difference between the shape character of the target object image and the shape character of the template image according to the coordinates of the plurality of discrete points of the shape character of the target object image and the coordinates of the plurality of discrete points of the template image; and 
 obtaining, by at least one computer hardware unit, the similarity between the shape character of the target object image and the shape character of the template image according to the first difference and the second difference.